d543838cf44e8d9e35d4cf08eca98463

This model is a fine-tuned version of facebook/opt-1.3b on the contemmcm/cls_mmlu dataset. It achieves the following results on the evaluation set:

• Loss: 3.4949
• Data Size: 1.0
• Epoch Runtime: 88.8886
• Accuracy: 0.2613
• F1 Macro: 0.2585

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 5e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• total_train_batch_size: 32
• total_eval_batch_size: 32
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: constant
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Data Size	Epoch Runtime	Accuracy	F1 Macro
No log	0	0	1.5773	0	3.3157	0.2586	0.2486
No log	1	438	1.6039	0.0078	4.0401	0.2699	0.2464
No log	2	876	1.5080	0.0156	5.8856	0.2620	0.1991
No log	3	1314	1.6033	0.0312	9.2265	0.2547	0.1133
No log	4	1752	1.4703	0.0625	13.4936	0.2507	0.1339
0.0846	5	2190	1.4476	0.125	20.1368	0.2553	0.1257
0.1868	6	2628	1.4210	0.25	32.8071	0.25	0.1023
1.3996	7	3066	1.3885	0.5	53.1310	0.2593	0.1839
1.3919	8.0	3504	1.4170	1.0	94.1025	0.2660	0.1992
1.2216	9.0	3942	1.5834	1.0	90.9216	0.2553	0.2401
0.7605	10.0	4380	2.2359	1.0	91.5593	0.2633	0.2551
0.3766	11.0	4818	3.4949	1.0	88.8886	0.2613	0.2585

Framework versions

• Transformers 4.57.0
• Pytorch 2.8.0+cu128
• Datasets 4.3.0
• Tokenizers 0.22.1